EU AI Act Transparency Statement
Momntum builds AI agents that speak with people every day, on behalf of the businesses that use us. We think anyone on the other end of one of those conversations is entitled to know how it works. This is how we approach that.
Last updated July 2026
What we build
Momntum builds AI agents that talk to people. Our platform, Laila, powers conversational agents that answer calls, reply on WhatsApp and web chats, and book appointments on behalf of the businesses that use us. Our customers configure those agents with their own knowledge, their own rules and their own brand. The agent that speaks to you belongs to the business you are dealing with. We build and run the technology behind it.
What we build
Momntum builds AI agents that talk to people. Our platform, Laila, powers conversational agents that answer calls, reply on WhatsApp and web chats, and book appointments on behalf of the businesses that use us. Our customers configure those agents with their own knowledge, their own rules and their own brand. The agent that speaks to you belongs to the business you are dealing with. We build and run the technology behind it.
You always know you are talking to AI
We design our agents to say so. Not in the small print, but at the start of the conversation, before anything else happens. On a call, the agent identifies itself as an automated assistant at the opening. In messaging, the first message does the same. And if you ask at any point whether you are speaking to a person, the agent will tell you the truth and offer to put you through to one. This is built into the platform. Our customers cannot switch it off.
We design our agents to say so. Not in the small print, but at the start of the conversation, before anything else happens. On a call, the agent identifies itself as an automated assistant at the opening. In messaging, the first message does the same. And if you ask at any point whether you are speaking to a person, the agent will tell you the truth and offer to put you through to one. This is built into the platform. Our customers cannot switch it off.
A person is always reachable
An AI agent should never be a wall between you and a human being. Every deployment we run has a defined route to a person. Our customers set the rules for when that happens automatically, and you can ask for it at any time.
An AI agent should never be a wall between you and a human being. Every deployment we run has a defined route to a person. Our customers set the rules for when that happens automatically, and you can ask for it at any time.
We do not train on customer data
We do not use our customers' data, conversations or materials to train, fine tune or improve any model or service beyond the specific deployment they were provided for. Not our models, and not anyone else's. We contract with our model and infrastructure providers on terms that prohibit them from training on customer data, and we enable minimised data retention where a provider offers it.
We do not use our customers' data, conversations or materials to train, fine tune or improve any model or service beyond the specific deployment they were provided for. Not our models, and not anyone else's. We contract with our model and infrastructure providers on terms that prohibit them from training on customer data, and we enable minimised data retention where a provider offers it.
Where data is processed
We process personal data within the European Union by default, and we use European regions where our providers offer them. Where a transfer outside the European Economic Area is unavoidable, we use a lawful transfer mechanism. We act as a data processor for our customers, who remain the data controller. We process personal data only on their documented instructions, under a data processing agreement that meets Article 28 of the GDPR. We maintain a list of the subprocessors we use, and we make it available to customers on request.
We process personal data within the European Union by default, and we use European regions where our providers offer them. Where a transfer outside the European Economic Area is unavoidable, we use a lawful transfer mechanism. We act as a data processor for our customers, who remain the data controller. We process personal data only on their documented instructions, under a data processing agreement that meets Article 28 of the GDPR. We maintain a list of the subprocessors we use, and we make it available to customers on request.
What our customers own
Our customers own their data and their content. That includes their conversations, transcripts, recordings and analytics, and it includes the knowledge, scripts, rules and conversation design that define how their agent behaves. They can take it with them. We provide a complete export in a documented, machine readable format, on request during the relationship and on exit, so that a customer is never locked in by their own material sitting in our system.
Our customers own their data and their content. That includes their conversations, transcripts, recordings and analytics, and it includes the knowledge, scripts, rules and conversation design that define how their agent behaves. They can take it with them. We provide a complete export in a documented, machine readable format, on request during the relationship and on exit, so that a customer is never locked in by their own material sitting in our system.
Who is responsible for what
We are clear about this because it matters. We build and operate the system. We are responsible for how it works, for the AI disclosure, for security, and for the platform behaving as specified. Our customers own and are responsible for the subject matter their agent works with, including any specialist, clinical, regulated or professional content. We do not author that content and we are not qualified to validate it.
We are clear about this because it matters. We build and operate the system. We are responsible for how it works, for the AI disclosure, for security, and for the platform behaving as specified. Our customers own and are responsible for the subject matter their agent works with, including any specialist, clinical, regulated or professional content. We do not author that content and we are not qualified to validate it.
What our agents do not do
Our agents are communication and scheduling tools. They are not decision makers. We do not build agents to:
Our agents are communication and scheduling tools. They are not decision makers. We do not build agents to:
- give medical, clinical, legal or financial advice, or to replace professional judgment
- make decisions that produce legal or similarly significant effects for a person, such as decisions on employment, credit or eligibility
- perform emotion recognition, biometric categorisation or biometric identification
- carry out any practice prohibited under Article 5 of the EU AI Act
How we govern this
We maintain an internal AI governance policy covering every system we provide and every AI tool we use ourselves. It sets out how we classify risk before anything goes live, how we design transparency in, how we handle incidents, and how we keep our team trained. We assess the risk classification of each system under the EU AI Act before it is put into service, and we reassess it whenever a system, its purpose or the law changes. Where a classification is uncertain, we take external advice rather than assume.
We maintain an internal AI governance policy covering every system we provide and every AI tool we use ourselves. It sets out how we classify risk before anything goes live, how we design transparency in, how we handle incidents, and how we keep our team trained. We assess the risk classification of each system under the EU AI Act before it is put into service, and we reassess it whenever a system, its purpose or the law changes. Where a classification is uncertain, we take external advice rather than assume.
We review the policy at least annually, and it is owned at executive level.
Raising a concern
If you have spoken to one of our agents and something was wrong, unclear, or made you uncomfortable, we want to know. Contact us at compliance@momntum.com and we will look into it. If your concern is about the business whose agent you spoke to, rather than the technology, we will pass it to them and tell you that we have. If you are a customer, your usual contact is the fastest route, and our contracts set out how we handle and report incidents.
If you have spoken to one of our agents and something was wrong, unclear, or made you uncomfortable, we want to know. Contact us at compliance@momntum.com and we will look into it. If your concern is about the business whose agent you spoke to, rather than the technology, we will pass it to them and tell you that we have. If you are a customer, your usual contact is the fastest route, and our contracts set out how we handle and report incidents.
Momntum AI Limited, registered in Ireland number 783108.
Upperdeck Office 209, Ferry Terminal, Dún Laoghaire, Co. Dublin, A96 X7F2.
This statement is reviewed at least annually.
Upperdeck Office 209, Ferry Terminal, Dún Laoghaire, Co. Dublin, A96 X7F2.
This statement is reviewed at least annually.
